NICORAEPOA (POACEAE, Robert J. Soreng2 and Lynn J. Gillespie3 POEAE), A NEW SOUTH AMERICAN GENUS BASED ON POA SUBG. ANDINAE, AND EMENDATION OF POA SECT. PARODIOCHLOA OF THE SUB-ANTARCTIC ISLANDS1 ABSTRACT Poa subg. Andinae Nicora (Poaceae, tribe Poeae) is raised to the rank of genus, here renamed as Nicoraepoa Soreng & L. J. Gillespie. Six speeific and one infraspecific combinations are made: N. andina (Trin.) Soreng & L. J. Gillespie, N. chonotica (Phil.) Soreng & L. J. Gillespie, N. erinacea (Speg.) Soreng & L. J. Gillespie, N. pugionifolia (Speg.) Soreng & L. J. Gillespie, N. robusta (Steud.) Soreng & L. J. Gillespie, N. subenervis (Hack.) Soreng & L. J. Gillespie, and N. subenervis subsp. spegazziniana (Nicora) Soreng & L. J. Gillespie. Lectotypes are designated for N. pugionifolia and N. subenervis. Taxonomy and nomenclature of similar Southern Hemisphere island species of Poa L. and other related genera are discussed. Poa sect. Parodiochloa (C. E. Hubb.) Soreng is emended to include, in addition to P. Jlabellata (Lam.) Raspail, P. cookii (Hook, f.) Hook, f., P.foliosa (Hook, f.) Hook, f., P. hamiltoni Kirk, P. ramosissima Hook, f., and P. tennantiana Petrie. The one species of Tzvelevia E. B. Alexeev is reunited with Poa, and the genus is transferred to Poa sect. Tzvelevia (E. B. Alexeev) Soreng & L. J. Gillespie. Morphological and anatomical characteristics of Nicoraepoa, Poa sect. Parodiochloa, and similar genera in Poinae are compared. Leaf blade morphology and anatomy were found to be particularly useful in characterizing Nicoraepoa and related taxa. A key is provided to distinguish Nicoraepoa, Poa sect. Parodiochloa, and morphologically similar perennial genera of Poinae. Key words: Austrofesluca, Bellardiochloa, Festucella, Hookerochloa, Nicoraepoa, Parodiochloa, Poa, Poaceae, Poinae, Tzvelevia. Poa subgen. Andinae Nicora (Poaceae, tribe Poeae) two are perfect-flowered. The lead author thought that was described for a group of robust, tough, gynodioe- the subgenus probably needed to be removed from cious species from Patagonia having stolons or Poa L. since he first saw material of the group in the rhizomes, long staminodes (1•3 mm) in pistillate early 1980s. With new material from collection trips plants, smooth or scabrous lemmas with glabrous or to Chile in 2000-2001 (by R. and N. Soreng) and pilose keel and nerves, and a callus that is glabrous or Argentina in 2001 (by P. M. Peterson, R. J. Soreng, N. with rigid hairs to 3 mm long (Nicora, 1977). Some Refulio, and M. Belgrano), the opportunity to study species also have awns at the lemma apex (e.g., P. types and other specimens, and the continued study of andina Trin.; Nicora, 1978; Soreng, pers. obs.). Nicora the large genus Poa (ca. 500+ species), the morpho- (1978) included six species in Poa subgen. Andinae logical and anatomical distinctness of Poa subgen. (P. andina Trin., P. borchersii Phil. (= P. chonotica Andinae has become clearer, supporting the conclu- Phil.), P. erinacea Speg., P. pugionifolia Speg., P. sion that most species of the subgenus and P. robusta Steud., P. stepparia Nicora), but retained the subenervis (placed in Poa subgen. Poa by Nicora, morphologically similar P. jlabellata (Lam.) Raspail 1978) do not belong in Poa. However, there was and P. subenervis Hack, in subgenus Poa as the latter always uncertainty about whether the group repre- 1 At the Smithsonian Institution we thank Scott D. Whittaker for assistance with SEM work, Paul Peterson for a collection trip to Patagonia, Dan Nicolson for answering many nomenclatural questions, and the Department of Botany for resources and funds for collecting trips. We thank the curators of BAA, CONC, K, MO, SGO, and SI for loans. We thank Annie Arehambault for DNA lab work at the Canadian Museum of Nature and the Research Division for providing financial assistance. We thank Surrey Jacobs for leaf material of Festucella and Hookerochloa, Stoney Wright for leaf material of Poa Jlabellata, and R. C. Gardner, J. Keeling, P. J. de Lange, S. D. Wright, and E. K. Cameron for an unpublished GenBank ITS sequence of Poa ramosissima. We thank Renee Fortunato and Instituto Nacional de Tecnologia Agropecuraria (INTA) for providing the scans of seven Flora Patagonica illustrations and permission to reprint them. 2 Department of Botany, National Museum of Natural History, Smithsonian Institution, Washington D.C. 20013-7012, U.S.A. [email protected]. 3 Canadian Museum of Nature, P.O. Box 3443, Station D, Ottawa, Ontario, KIP 6P4, Canada. [email protected]. ANN. MISSOURI BOT. GARD. 94: 821-849. PUBLISHED ON 13 DECEMBER 2007. 822 Annals of the Missouri Botanical Garden sented a new genus or could logically be placed within of Poa (Gillespie & Soreng, 2005). Parodiochloa, an existing genus in subtribe Poinae and, if it was which included the single species Parodiochloa distinct, how best to delimit it. flabellata (Lam.) C. E. Hubb., was also shown to Several genera of Poinae had to be considered in belong within Poa and is now placed in its own the decision on how to treat Poa subgen. Andinae. section within Poa (Soreng et al, 2003a; Gillespie & Understanding the relationship of P. flabellata (sect. Soreng, 2005; Gillespie et al., 2007). Resolution of the Parodiochloa (C. E. Hubb.) Soreng) to P. subgen. generic taxonomy problem remained incomplete, Andinae species was critical in deciding what genus however, as there is also a small group of sub- name to use, because that species had been trans- Antarctic island species of Poa that were thought to ferred to a new genus, Parodiochloa C. E. Hubb. be closely related to Parodiochloa (Edgar, 1986), and (Hubbard, 1981). These groups share morphological thus potentially related to Poa subgen. Andinae, and features such as thick-ribbed leaf blade adaxial no molecular data were available for these or for the surfaces and awned lemma apices. Other genera monotypic sub-Antarctic island genus Tzvelevia. morphologically similar to Poa subgen. Andinae, or Another genus of subtribe Poinae, the Eurasian genus otherwise putatively allied to Poa (Alexeev, 1976, Bellardiochloa, was previously placed in Poa (Clayton 1985; Garcia-Gonzalez, 1983; Clayton & Renvoize, & Renvoize, 1986) or was thought to be closely related 1986), are Austrofestuca (Tzvelev) E. B. Alexeev s. (Garcia-Gonzalez, 1983). Molecular analyses have str., Bellardiochloa Chiov., Festucella E. B. Alexeev, consistently shown that Bellardiochloa does not Hookerochloa E. B. Alexeev, and Tzvelevia E. B. belong within Poa (Soreng et al., 1990; Davis et al., Alexeev. As presently circumscribed, these genera 1993; Soreng & Davis, 2000), but it has only recently include a total of only nine or 10 species that all, been demonstrated that this genus is not a sister to or except for Bellardiochloa of southeastern Europe and direct ancestor of either Poa or Poa subgen. Andinae the Middle East, occur in the Southern Hemisphere, (Soreng et al., 2007; Gillespie et al., 2007). in Australia, New Zealand, Tierra del Fuego, and on The present paper outlines the nomenclature and sub-Antarctic islands. All of the above genera have taxonomy of a new genus established here to been controversial, and, except for Bellardiochloa, all accommodate Poa subgen. Andinae. We discuss how their species were originally described in Festuca L. it may be distinguished from Poa and other genera of or other genera other than Poa and have since been Poeae subtribe Poinae, especially those genera placed in the above new genera and, in most cases, thought to be most closely related. A revised taxonomy also in Poa (Hooker, 1844-1860; Alexeev, 1976, is also presented for the morphologically similar Poa 1985; Clayton & Renvoize, 1986; Tzvelev, 1989; sect. Parodiochloa and Tzvelevia, which here is Jacobs, 1990; Watson & Dallwitz, 1992; Sharp & reduced to a section in Poa. New data on leaf blade Simon, 2002; Soreng et al., 2003a; Gillespie & structural characters are presented and discussed Soreng, 2005). Poa flabellata, for example, had also here. Gillespie et al. (in prep.) will provide details and been placed in Dactylis L., Festuca, Parodiochloa, discussion of our new molecular analyses, including and Sesleria Scop. both plastid and nuclear DNA sequences, which Recent chloroplast DNA (cpDNA) analyses placed support the taxonomic conclusions of this paper. species of Poa subgen. Andinae outside of Poa in a clade with (Soreng et al., 2007) or near (Gillespie & MATERIALS AND METHODS Soreng, 2005; Gillespie et al., 2007) the panarctic genus Arctagrostis Griseb., while ITS data support Specimens, including types, were borrowed from a clade of these two taxa plus two Australian genera, the following herbaria: BAA, CONC, K, MO, SGO, Festucella and Hookerochloa (Gillespie et al., in and SI. Additional material was reviewed at BAB, prep.). Hunter et al. (2004) provided the first HIP, LE, LP, SI, US, and W. Specimens at US were molecular evidence that the latter two genera, placed used for anatomical work. Additional representative in Austrofestuca by Jacobs (1990), were closely related specimens are cited as vouchers for geographic ranges to each other and isolated from Poa, and that of the species cited beyond those ranges vouchered in Austrofestuca s. str. could be included within Poa. selected published works. Vouchers for taxa depicted The latter result was not unexpected as A. littoralis in Figures 1•3 are given in the figure legends. Leaf (Labill.) E. B. Alexeev is morphologically quite surfaces were examined and photographed using similar to P. macrantha Vasey. Additional molecular a Philips XL30 ESEM LaB6 microscope (Eindhoven, studies (Gillespie & Soreng, 2005; Gillespie et al., The Netherlands) at the National Museum of Natural 2007; Soreng et al., 2007) and morphological History, Smithsonian Institution. Sections of blades comparisons have corroborated this result, and were taken from about 1 cm above the collar of lower Austrofestuca is now treated as a monotypic section culm leaves of herbarium specimens, attached to Volume 94, Number 4 Soreng & Gillespie 823 2007 Nicoraepoa in South America stubs, and sputter coated with 12-18 nm of 60:40 several times wider (Fig.